technical noise measurement question

[Speedskater]

Digital bits don't have a voltage value until you run them through an D to A converter.

[jackinnj]

Originally posted by Kevin Graf
Digital bits don't have a voltage value until you run them through an D to A converter.

Read the prospectus carefully before you invest.

[poobah]

You're almost correct about bits and voltgage values. Amp-ohms are the proper unit in this situation.

[EC8010]

Originally posted by jackinnj
Bit 24 is 59.6nV, bit 23 is 119.2nV -- that's a pretty big gulp of water.

OK, I worked backwards, and you're referring to 2V. If I've done my calculations correctly, 59.6nV is the self-noise produced (over 20kHz) by a perfect 10 Ohm resistor. I stand by my original observation that on most 24 bit ADCs, the four least significant bits are almost certainly rubbish.

[poobah]

You did mean 59.6 nanoAmp-Ohms... right?

[jackinnj]

Wasn't that the point I was making as well ? there is no resolving power down in the nano's, and it's inadequate until you get over a microvolt or so.

[EC8010]

Originally posted by jackinnj
Wasn't that the point I was making as well ?

Ah, well I wasn't sure if I had understood your colloquialism correctly.

As for the nanoamp-Ohms, there are applications where the noise is given in numbers of electrons.

[jackinnj]

Originally posted by EC8010

a perfect 10 Ohm resistor.

One of the things encouraged me early on with the HP 3581 and noise measurement -- it will run off batteries and with its 3Hz crystal filter bandwidth you can compare the theoretical noise and get a good idea of whether your nanos are really nanos.

To acurately measure noise with a DVM you need a darned good outboard RMS detector as well. The venerable AD536 is so-so for this purpose. Fortunately, the push for energy efficiency and power measurement has resulted in a windfall for us DIYr's --

[EC8010]

Originally posted by jackinnj
To accurately measure noise with a DVM you need a darned good outboard RMS detector as well.

I've never really liked noise measurements presented on a DVM, whether it's an external bodgebox to your Fluke or the internal bits on an Audio Precision. Give me an analogue meter with properly designed ballistics (a PPM) so that I can watch to see if the meter twitches randomly or rhythmically...

[jackinnj]

Certainly when you are nulling an LC circuit the old d'Arsonval is the way to go -- with noise you should be integrating the RMS value over a period of time -- if you have a scope which allows you to see an envelope it would be as helpful.

I should make an animated GIF of the meter on the HP3581 -- it gives me a headache.